To close out day one of our inaugural KBW AI Infrastructure Summit, I am delighted to welcome Ben Gagnon, CEO, and Jonathan Mir, CFO of Keel Infrastructure. Ben and Jonathan, thank you both for joining us. Thank you for having us. Good morning. Ben and Jonathan, I have started each fireside with the same sentiment question. I would like to get Keel's perspective as well. Sentiment toward miners moving in AI infrastructure has become, in my view, more skeptical since June, because the infrastructure layer is tied to the health of the entire AI stack. Investors are looking at everything above it, from deployment and safety concerns to token optimization, pressure on the frontier labs, political opposition, higher yields, and circular financing. When you look across that wall of worry, what do you think the market is getting right? What is it getting wrong? What changes the sentiment positive from here? Yeah. It is a great question. A little bit loaded, but I will try and address it as best as I can. I think the big one that people are focusing right now is on the political side of things and the upcoming midterms in November. There seems to be a lot of expectations being baked into what happens in the November midterms and how that has an impact on the industry, overall, as well as the credit markets and the cost of capital for all of these developments. I think the challenge is, with a lot of things, is that usually the loudest voice in the room is often one of the most misinformed. What we see as one of the challenges in the political conversation is that a lot of the points against data centers and against these kind of developments are really at odds with the development project facts themselves. What we see over and over again are the same concerns around water usage, around electricity rates, property rates, all these different things that people are focusing in on. I think most of the developers who are taking that long-term view that Keel has, that we are investing in the community for decades, not for months or years or anything like that, decades, that there is a real good way to line things up for both sides that is a mutual win. That takes some time to get people comfortable with and understand. I think most of the concerns you see from communities are proactively addressed by the good developers like Keel. We are very good about going into the communities early, transparently, and addressing questions before we have ever filed for a permit. I think that is one of the big separations between a company like Keel and maybe a speculator or a wildcatter who is going out there saying, "Hey, I have had this property for 20 years, and clearly there is more value potentially in this property now than anything else. How do I get something out of it as quickly as possible?" I think there is a big divide there. In the communities that we operate in, we have quite a lot of community support from the key leaders in the communities, and we have also been able to win over a lot of folks who were very skeptical about the projects and about Keel specifically coming into the town. I think that one is a bit overstated, and I think unfortunately, until we get to a number of these assets turning online and seeing actually how they work and seeing how the community is impacted or not impacted, it is going to be a while before the sentiment changes, at least on the community side of things. But the more of these developments that come in place, the more of the economic benefits are realized, and more of the concerns are assuaged. I think they are going to be in a better spot on the sentiment side of things. Yeah. I think that— Maybe— Sorry, go ahead, Ben. Maybe Jonathan has a comment on the capital as well. Because I think that's probably, in addition to the political, would be the second most important thing. Yeah, certainly. I think as a result of very significant supply from a very concentrated group of issuers, there's a stacking of risks that have taken place. There are lots of capital-intensive industries. There aren't many where five or six issuers essentially underlie the market. So you have industry concentration, general backup in interest rates, some concern about, as Ben has described, the fundamental folks who are driving demand in the space. In a market where the Fed is trying to back up rates, that's just served to both widen the cost of debt financing, and if one thinks about things from a fundamental basis, if risk-free rate rises, so should one's cost of equity. So we're just in a period where investors are applying higher discount rates both to the debt and the equity as they value companies. Market conditions change over time. What's good for folks in our role is that the markets are functional, and people can get financing done, although it's more expensive than it was six months ago. Debt investors are focused on quality of structure and quality of lease, but certainly things have gotten more expensive. There'll be a cycle to that. It was great to hear both of your perspectives on that. I want to bring the discussion down to what you're seeing in the leasing market. Demand still appears very strong, but there's clearly more competition from other miners, from established private data center developers, new entrants. What factors are winning a lease today, in your view? Across Panther Creek, Sharon, and Moses Lake, where do you think Keel is most differentiated? Yeah. It's a great question. I think the clear driver for what gets a lease done on the tenant side is timeline to power, timeline to power, timeline to power. That's been the prevailing narrative, and frankly, everything else that you mentioned is feeding into that and compounding the importance of timeline to power, right? The more there is political conversation focused around HPC and AI, the more timeline to power matters, the more there's questions around what's going on with the credit market and the cost of capital over time, the more timeline to power matters. It really is the primary factor, and I think we've even seen that in the last couple of weeks override the scale question, which was previously right there along with timeline to power with scale. Now you have everyone looking at how do I get sub 50 MW sites because the focus is so much on that timeline to power, they're going to go secondary on all the other concerns. That's really what gets it done, I think on the tenant side of things. I think what separates us from most of the other developers is we didn't focus on low barrier to entry markets. We specifically focused on high barrier to entry markets. The standard that we applied for ourselves and the standard that we were holding ourselves to very early on was much higher than it was for a lot of other developers and in a lot of other markets. I think what you see in Pennsylvania was that the executive order that came out a couple of months ago really was catching up to the standards that Keel was already holding itself to, as opposed to Keel needing to adjust to the changing political environment. I think we've really positioned ourselves well in that regard, and I think that helps to give us a head start over all the other developers who are trying to adjust to the political reality in real time. We held ourselves to a higher standard, and we're just seeing the political reality, I think, trying to catch up to how we view responsible development. I'd just add that our strategy of emphasizing liquidity on the balance sheet, we had, I think, almost $825 million of liquidity as of last quarter when we reported earnings, and it's quite close to that right now. That helps in discussions with potential customers and our potential partners around construction and otherwise because they know Keel is going to be there to stand by its commitments. We think that strategy has been really helpful to us as a first-time developer. Thank you. Turning from positioning to execution, on the Q2 call you highlighted some remaining permitting work at Panther Creek and Sharon, while Moses Lake was expected to be your first site energized in 2027. To the extent you can say, what has moved since then? What still needs to happen for each site to reach NTP? Yeah, happy to address that. First and foremost, I think nothing has really gotten structurally harder since the Q2 call. If anything, we've continued to progress on the permitting progress across Panther Creek, Sharon, and Moses Lake. All of those have been moving forward. We did have a little bit of a slowdown at Panther Creek by a couple of months, but that is starting to get resolved, and we're continuing to see good progress there. Moses Lake is already under development, and the remainder of the permits that are sitting there at Sharon and Panther Creek are really engineering work focused around wastewater runoff, sewage connections, these sorts of things which are very not politically contested, right? They're just fundamentally engineering questions. They're have you done the work? Does it comply? Does it meet with our expectations? They're not things where somebody's going to throw a protest over that kind of a permit. We've secured all the key ones, and the ones that are remaining are quite minor engineering works that are well progressed. All of that is progressing more or less on schedule. The executive order probably did slow things down for a little bit in Pennsylvania, but I think we are, at this point, we're right back on track and proceeding on schedule. Very excited about the developments that we've had, especially breaking ground at Moses Lake. That's a very exciting progress for us and hopefully we'll be there soon with the other projects. Congrats, by the way, on breaking ground there. Thank you. You're negotiating, Ben, you've said with several potential customers across the portfolio at the same time. What differences are you seeing in pricing power across, whether it's hyperscalers, chip companies, AI labs, neoclouds? Is there a customer core that you think provides the most alpha or best risk-adjusted returns today for your lease portfolio, and why? You can't provide a blanket answer, unfortunately, to that. You can't just say hyperscalers are always going to provide the best returns across our portfolio. Really, the reality is that the demand from the customer is what is determining their willingness to pay. What we've seen is that the more the tenants have invested into GPUs and compute that they can't deploy, the more aggressive they are going to be in the negotiations and pricing in order to secure the infrastructure necessary to monetize those investments. This has actually, I think, been a very consistent investment thesis from Keel from the beginning, even before we had gone through the rebrand to Keel, is that our investment thesis was that people are investing hundreds of billions of dollars into GPUs and compute, that there is not the infrastructure to deploy. The emphasis is going to be around, h ow do you deploy with the fastest timeline possible, the least amount of risk, and start generating revenue as quickly as possible? Because the alternative is making hundreds of billions of dollars into compute that's going to sit in a warehouse somewhere, depreciating in real time, not generating revenue and losing market share to your competitors. I think that dynamic has continued to prevail. If anything, the macro over the last six months has really driven that to a much more extreme level because of all the project delays, because of the cancellations, because of the setbacks. People are getting increasingly more eager to deploy capital and to secure that infrastructure with increasingly fewer options at the table. I think, one of the things that we've seen is how people are adjusting to that dynamic, going to smaller sites, going behind the meter, whatever they can do to try and accelerate that. I think they're going to get really creative to do so. Jonathan, you mentioned your indicator, obviously the rise in the 10 year, I think it's up some 80 basis points from Q2. V- spec debt that we've seen in the market has moved with it. An environment of maybe increasing rates or higher for longer. Can you pass the higher cost of capital through to customers in terms of higher yields on cost on the leases that you're potentially negotiating? If so, what are you seeing in those negotiations and customer willingness to accept higher lease rates? Yeah. I'd say customers understand a couple of points from the broader capital markets, which is one, the absolute increase in financing costs. That's just math. There's nothing much to debate over on that front. As well as the need for a simple and straightforward lease structure, including an investment grade or investment grade wrap, because the distinction in financing costs between the two has become so extraordinarily wide. The difference between a 7 and 7/8 coupon on a Meta backed lease or 12%, 13- odd% on something with a lesser credit quality backing. That certainly becomes relevant in discussions, and helps us establish some boundary conditions in terms of what works and what doesn't work for us from a leasing perspective. Okay, then maybe just comment too on, with spreads widening, maybe just more holistically how the funding environment has changed since, I think in August you said that the project debt markets hold enough depth even as spreads widened. How far does your current liquidity take you if maybe leasing were to take longer than expected? You mentioned the strong liquidity position Keel has, and so forth. Yeah. I'd say first, we continue to think that the project finance markets remain functional and reasonably constructive, albeit more expensive than they were in Q2. We're comfortable with our ability to finance debt in the amounts needed for projects like Sharon or Panther Creek. We have great financial advisors, and they're comfortable with the quantum of debt that we would need to issue around either of those projects. So that's a really solid foundation for us. In terms of liquidity, our strategy, as has been discussed, is to ensure that we're fully funded for all of our SG&A at least through 2028 and well beyond. That remains the case. At the same time, we're able to focus on some critical growth capital opportunities in front of us. I'd say our liquidity position is in at least a good as position as it was in Q2, and we've taken some actions to give ourselves more flexibility and more runway. The hashprice has gone up too, since Q2. We don't talk about that much anymore in this space, but that's been a positive. Ben, on the Q2 call, we talked about this earlier, but you described Pennsylvania as a unique centrist state. Since then, the Governor has signed the executive order you spoke about in August that establishes regulatory guardrails and a new permitting framework for data center projects. PJM, and I think Pennsylvania's PUC both want new data centers curtailed first in great emergencies unless they bring their own capacity. Correct me if I'm wrong on that. They want maybe data centers to cover PJM's backstop costs. So how do these rules apply to Panther Creek and Sharon, and how are they shaping your lease negotiations? Yeah. The executive order that came out in August, as I mentioned earlier, was largely, I think, catching everyone else in Pennsylvania up to the standards that we were already approaching. I think the big ones that people focus in are funding the infrastructure and making sure that local rate payers are not footing the bill for us to build out the transmission infrastructure and the additional generation capacity for these sites. That was already something that we were signed up to do across all of our projects. There's a lot of areas here where the headline might seem quite scary, but the actual reality of the implementation is very different than sometimes what the headline implies. The reality is that, for the most part, our sites were well advanced enough. We had secured the community approvals that the executive order was really designed to try and enforce people to get in advance of submitting all of these projects. I think what it really did more than anything was cement our moat in place that we've described so frequently. It's going to be very, very hard for other companies to come into PA and to adjust to the standards that we've been holding ourselves to for the last couple of years. If you're used to operating in an environment like Texas, where the permitting is a three week to a three month process, and then you want to come into Pennsylvania, it's a completely different ballgame. These are not apple and orange comparison. This is like, I don't know, a bowling ball and a grape kind of comparison. They're very, very fundamentally different things. I think, even when you look back to the innovation summit that happened last year with Governor Shapiro and President Trump and Senator McCormick and others really talking about how they're embracing data centers as an industry, and they're excited about the economic development and reinvigorating this part of America, which is one of the most economically depressed parts of America, and hasn't seen these levels of industrial investment in decades, if not centuries, is a real important thing. They announced 60 GW of development, and I think we're farther along than even those 60 GW of development that were announced with all of that political support. I think that's another example about how we were holding ourselves up to a higher standard because we thought that was the right thing to do, and now everyone's kind of catching up to us. That is something that is noticed by the tenants as well. The tenants who we are negotiating with and are in very detailed conversations with, they obviously are reacting and trying to understand the political dynamics, and in many situations are very comforted by the fact that we have addressed these things prematurely and in advance of any sort of requirement to do so. I think that sets us up really, really well, and we believe that the sites that we have in Pennsylvania have just become that much more scarce and potentially that much more valuable as a result. I think this sets up a good. I am going to jump a little to a cap rate question for you both, but I think what you just said sets it up well before I go into some other questioning. This has been a kind of big topic I have been having of conversation with investors, is just like as rates are moving up, I think people are sort of all over the place in terms of where they may be valuing leases. We do not have cash flow online right now. There is not EBITDA online, so we do not really have any private market benchmarks to sort of point to or not ones that happened recently, so to speak. So I guess, given what you have just said around barriers to entry in Pennsylvania being stronger, the scarcity of your assets, maybe talk to how you think higher rates affect the value and sort of implied cap rates of your prospective leases on those three sites, Sharon and Panther Creek, and then maybe Moses Lake as well in that environment. The first thing to address here is what I previously mentioned with the opportunity cost of if you have already made these sunk investments into the compute itself, which represents the vast majority of the investment into growing your compute stack, you are much more incentivized to pay more and higher and higher prices in order to deploy faster with greater certainty. If you think about it just kind of in simplified terms, let us just say using round numbers the infrastructure spend is $10 million a megawatt. The compute spend is $40 million a megawatt, and it is likely a lot higher, but simple numbers. What you are looking at is the infrastructure expense is 20% of the total expense to actually grow your compute stack. Is it worth you to argue over basis points for the project overall when you really need to deploy that and you've got 80% of your cost sunk into compute? I think the reality of the situation is that, certainly people are always trying to get the best deal possible, but the best deal is often the most de-risked, the fastest timeline to power, the most secure. I think there is a strong economic incentive in place across the market to pay a higher rate in order to secure the power you need to monetize your investments that you've already made. I think that dynamic only gets stronger by the day, and only gets more valuable for assets like Panther Creek and Sharon. Thank you. I want to switch. I think Keel has a unique position in terms of the design with Panther Creek and Sharon. I think there's a Vera Rubin focus. I am curious to get your thoughts, Ben and Jonathan, on with NVIDIA's rack architecture still evolving, when do you expect to lock in a basis design for those sites where you have high confidence in the build costs on a CapEx per megawatt basis? Maybe you are already there from what you said on the last call, but maybe you can talk to that and what you are seeing there. Yeah. So there are two things. Well, there are a few things there, Stephen. First of all, obviously, we as a developer have our own basis of designs for the sites. Then the tenants also have their own basis of designs, and it is about how do you find the workable solution for both parties. Maybe it is a modification to our design, maybe it is a compromise on what they actually wanted. But the challenge with Vera Rubin is not so much the math, it is the network of infrastructure and equipment suppliers around it, in order to help support those basis of design decisions. For example, there are not 800 V DC infrastructure providers to handle the cabling for the distribution throughout the facility. The breakers are completely different requirements. All of these things actually need to be are actually being solved by the industry in real time. When it comes to the basis of designs, they are actually changing as the industry is evolving, and there is no standard, and this is the one way that everyone is going to build out Vera Rubin infrastructure, because we are talking about a fundamental shift in electrical distribution, in heating or in cooling, as well as the communication, even just in a rack. So we have our own basis of designs. It is based off of the reference designs from NVIDIA and our architecture and engineering partners. The customers all have their own basis of designs, and what I could tell you is from the Vera Rubin basis of designs that we have seen from the various tenants, there are quite a few different approaches to how to make this work and how to make it cost-effective. We'll end up going with whatever is the workable solution between us and that tenant for the Vera Rubins. The good thing for us is we started on this. I think we started talking about building out Vera Rubin infrastructure before anyone else in the industry was talking about building out Vera Rubin. We've anticipated this for a long time, and we've been working towards it in anticipation. Okay. Thank you. That's really helpful. Turning to the longer-dated pipeline, maybe start with Sherbrooke. On the Q2 call, you said the city and utility approvals were complete, with only the provincial minister's sign-off remaining to consolidate the three existing power agreements into a single 96 MW HPC agreement. Where does that approval stand today? Once it is received, what maybe permitting or commercial work remains before you can formally begin or bring that site to market? Yeah. It's still waiting on a signature from the minister. The way that it works in Quebec is they have effectively kind of like an economic tsar who has a sign-off authority on any industrial consumption of electricity over 5 MW. That's a relatively low bar. Every industry, every even small business who's operating over that 5 MW needs to go through the same path with the minister. We're working through that. We don't have a clear indication as to when that would be resolved. There is no obligation on the minister's part to respond within X period of time. It's really just a matter of us working with the minister and with their department to try and get all of that done as quickly as possible. The good thing about Quebec, beyond getting that approval, which we do still anticipate is coming our way in whatever period that we can forecast, is that permitting is very, very simple in Quebec relative to where it is in Pennsylvania. This is an area where the difficulty is on getting the power. This is where Amazon's been rejected, where Microsoft's been rejected, where all these companies have been rejected on power. That's the hard thing that we have solved 99% of for just with the one signature remaining. Getting the construction permits is measured in weeks, not quarters or years. That's a relatively simple process, and we've already obviously have our own basis of design for the Sherbrooke site that we'd be looking to base things off of. As soon as we get that approval, we are going to be ready to go, and we are going to be ready to go with, again, energy that is very, very difficult to source. It is one thing to say that energy is scarce overall. But the markets that we are operating in, Pennsylvania, Washington, and Quebec specifically, are some of the highest demand markets with some of the most scarce opportunities to grow. What we would expect, I think, in Canada overall, but especially in Quebec, is you should anticipate higher overall rates that you should be able to extract out of customers than you would megawatt for megawatt in the United States. You should expect better PUEs than you do in the United States because of the dry, cooler climates. You should expect kind of overall, on a megawatt per megawatt basis, probably better project yields than you would in the United States. But that is because of the difficulty in getting all those approvals that we have worked so hard and advanced so much on. We think that is actually a very interesting opportunity for us in Quebec because we know the demand is there. We know the demand is there from some very high-quality potential Canadian customers who are looking to fulfill their data sovereignty requirements and have really no ability to grow outside of opportunities like ours. On Scrubgrass, I think, and we have a few minutes left, but I wanted to get a couple more questions in hopefully. You are still working through a 750 MW utility load study and the engineering for roughly, I think, 550 MW of potential on-site generation. What has progressed since the Q2 call, and when do you expect enough power and timing certainty to begin maybe formal tenant discussions? Could the grid connected and behind the meter be developed in parallel with each other? Yeah. A couple of different questions there. First, the detailed load study for the grid connection is progressing well. I think we originally outlined that we would hope to have an update for the market around end of the year or January, and that is still kind of the timeline that we should expect to have an update on the detailed load study. The good thing about the detailed load study and that particular location is just how close it is to the substation that we would be tying into. It is 3 mi from our site to the substation we would be tying into. At Panther Creek, we are building a line that is 30 mi. So we are talking about twice the power capacity for a line that is 1/10 the length. It is a relatively easier build. It is a relatively faster build and a cheaper build. To go 3 mi on an existing easement we already own and operate versus 30 mi off of easements that we all have to negotiate. With regards to the gas connection, still making very good progress on the lateral engineering and design process there. We hope to be providing an update to the investors around the same time that we are providing an update to investors on the detailed load study, so we're not giving them a piecemeal, "Hey, it's this, or it's that," but more of a comprehensive update as to, we anticipate this is a 1 GW plus campus, and this is how it breaks down across time and energy sources. When it comes to leasing, still far too early to be discussing leasing at a site like Scrubgrass. Frankly, we haven't even submitted permitting for sites like Scrubgrass because we're still working on securing the power. The sequence of events for us is secure the power, permitting, and engineering, followed by leasing. I'm going to ask you another bigger, higher level question, longer term question. I've asked this to others as well this morning, but once these projects are stabilized, your leases and stabilized leases, what's the long-term plan? Do you expect to hold Keel's interest in those for the cash flow or monetize maybe assets over time and recycle that capital into new projects? Let's start there. Maybe I'll give a quick answer, then I'll let Jonathan follow up with some more— Yeah. —detailed results. But at the end of the day, Stephen, what we're aiming for is what's going to drive the most shareholder value. I think that means that we want to focus on maximizing our NOI and maximizing our multiple, broadly speaking. That being said, we are a portfolio manager, and we're always going to be looking for opportunities to rotate the capital into new projects or maybe find better projects that are going to generate better yields and is a better move for us as a portfolio manager of these infrastructure projects. Jonathan? I'd echo Ben. We view ourselves as essentially a portfolio manager, as an owner and operator of a developer and owner of long-term infrastructure. We want to maximize long-term risk-adjusted ROEs across the biggest equity base possible. Part of being a good investor in infrastructure involves recycling capital so that there's equity for new growth investments. That's absolutely part of our business plan. I think investors should expect us to be thoughtful around making sure any capital we have at hand is being used in the most productive way possible. If we experience capital appreciation over time as a result of stabilization or other factors, there are a lot of ways we can finance, either at the asset level or at the parent level for growth, and we'll pursue those. Great. Ben, Jonathan, we're going to leave it there. Thank you both for a terrific discussion, and for helping us close out the first day of our inaugural KBW AI Infrastructure Summit. Thank you as well to all of today's speakers and everyone who joined us throughout the morning. We look forward to welcoming you back for day two tomorrow for TeraWulf, CleanSpark and MARA Holdings. Thank you. Stephen. Thank you. Thank you, everyone, for everything. Thank you.